Studies examining staff attitudes toward people with intellectual disability have traditionally used pre-determined categories and models or been open to researcher bias. The use of methods derived from personal construct psychology permits an objective investigation of staff views and attitudes without such limitations. Fourteen staff from an inpatient intellectual disability service were interviewed about their perceptions of clients with challenging behavior primarily using the repertory grid technique as developed from personal construct theory. Staff was found to construe their clients and their behaviors in a heterogeneous manner that was not readily reducible to a group average, and they did not make or use attributions about them in a consistent manner. Future research should incorporate work culture and the staff–client relationship. The results have implications for clinical decision making, team working, and clinical supervision.
Various researchers have proposed that the “challenging behavior” exhibited by people with intellectual disability may be unintentionally maintained by the responses of staff who are, in turn, part of a process of mutual social reinforcement (Emerson, 2001; Hastings & Remington, 1994; Oliver, Hall, & Murphy, 2005). Therefore, the decision of staff to engage or not in a particular response to a particular behavior can be crucial, which has lead researchers to use Weiner's attributional theory of helping behavior (1986) with its emphasis that staff beliefs about the causes of the challenging behavior would influence their emotional and behavioral responses to clients. To date, there is equivocal support for Weiner's model, with some studies finding results in the opposite direction to that predicted by the model (Bailey, Hare, Hatton, & Limb, 2006; Dagnan & Cairns, 2005; Dagnan, Trower, & Smith, 1998; Jones & Hastings, 2003; Lucas, Collins, & Langdon, 2009; McGuinness & Dagnan, 2001). Possible reasons for this lack of substantive evidence for Weiner's model were summarized by Willner & Smith (2008), who reported the widespread use of questionnaires and hypothetical vignettes, which called the ecological validity of such studies into question, as well as noting the scant evidence for the efficacy of carers' proffered “helping” and the limited attention paid to contextual and historical factors that may influence staff attributions. Many studies report an overwhelmingly positive response from staff about clients and their intention to help them, but the use of hypothetical scenarios may encourage social desirability response biases. Finally, there is the problem that such studies tend to focus on capturing an “average” response from participants.
An alternative approach to understanding human action, which explores the individual's perspective while minimizing interviewer bias, is the repertory grid technique. The technique was developed from personal construct theory (Fransella & Bannister, 1977; Kelly, 1955), which is predicated on the notion that people use their experiences to interpret the world, a process called construing. A person's unique constructs are used to make sense of the elements in their world, such as people, places, and hypothetical ideas such as the “ideal self.” Repertory grid techniques have been used in numerous studies of health professionals' perceptions of their work and their clients. Inpatient nursing staff have conceptualized the “ideal” or preferred patients as easier to care for and less impaired, using repertory grid techniques, although they may say they prefer the more complex clients in an interview (Barnes, 1990). Studies with staff in inpatient settings have shown that inpatient staff made more critical judgements about clients and perceived clients more negatively and different from themselves (Ralley, Allott, Hare, & Wittkowski, 2009; Winter, Baker, & Goggins, 1992; Woodrow, Fox, & Hare, 2011). Thus, one of the repertory grid technique's strengths is that it can simultaneously indicate inconsistencies in people's views and minimize social desirability biases as compared to using either questionnaire or interview methods. Of particular interest is the “cognitive complexity” of an individual's construct system (Bell, 2004), which is taken as indicating the degree of flexibility and fine discriminations between constructs within the system, allowing the individual to take different perspectives on that particular topic. Studies have suggested that low cognitive complexity of a repertory grid may be associated with a greater distinction between clients and nonclients (Blundell, Wittkowski, Wieck, & Hare, 2011; Ralley et al., 2009; Woodrow et al., 2011).
The current study aimed to extend previous findings from the attributional theory of helping behavior in this field by exploring how care staff in one intellectual disability service perceived clients with intellectual disability whose behaviors were challenging, using their own words. Broad concepts related to standard attributional dimensions (Stratton, Munton, Hanks, Heard, & Davidson, 1986) were included with the aim of understanding how these were construed by the participants, as questionnaire-based studies assume that respondents will construe the items in the same ways as the researchers, which may not be the case. The specific aims of the study were to explore how carers perceived: (a) clients whose behavior was thought to be caused by internal (client) versus external (environmental) factors, (b) clients whose behavior was thought to be under the client's control versus outside the client's control, (c) clients whose behavior was thought to be temporary versus permanent or enduring, and (d) clients versus carers.
Participants were recruited from an inpatient assessment and treatment unit for adults with intellectual disability. The seven-bedded unit provided 24-hr support from nurses, clinical support workers, and domestic staff, together with input from other health professionals. Typical reasons for referral to the service included placement breakdown because of behavior that challenges, medication reviews, and assessment of other health-related questions. The study was approved by a National Health Service Local Research Ethics Committee.
The participants were 14 out of a possible 16 clinical staff (71% female). Seven of these were learning disability nurses and seven were clinical support workers. The mean number of years of experience with people with intellectual disability and/or behavior that challenges was 17.21 years (SD = 9.1 years), ranging from less than 1 year to 31 years of experience. All except one of the participants used all 10 elements in their repertory grid generating between five and twelve constructs each.
The methods deployed in the current study were based on the study by Ralley et al. (2009), with a repertory grid being completed with each participant during an audio-recorded interview lasting 90–120 min. Participants were asked to think of people they had known well who fitted the following role titles: (a) a client whose challenging behavior has internal causes (personal to the client), (b) a client whose challenging behavior has external causes (in the client's environment), (c) a client in control of their challenging behavior, (d) a client who is not in control of their challenging behavior, (e) a client whose challenging behavior is permanent (enduring, unlikely to change), (f) a client whose challenging behavior is temporary, (g) a client without challenging behavior, (h) yourself, (i) a hypothetical ideal client, and (j) a carer who struggles. The triadic opposite method (Caputi & Reddy, 1999) was used to elicit bipolar constructs: The element cards were presented upside-down, and the participant selected three at random. The participant then thought of a way in which two of them were alike and different from the third, for example, “These two have lots of family contact.” Participants were then asked to provide the opposite of this elicited construct, for example, “having no significant others.” Participants then selected which was the most preferable pole of the construct (e.g., “It's preferable to have lots of family contact”) and rated all elements on this bipolar construct along a 7-point scale. The process was repeated until approximately 10 constructs had been generated or no new constructs could be obtained. All of the sessions were recorded, with participants being asked to clarify the meaning of their constructs. These “real time” quotes were subsequently used to complement the analysis of the individual repertory grids.
After the repertory grid had been completed, the participant was asked to define each of the elements and describe the behaviors of each person and their relevance. The aim of this was to further explore and check their construal of the attributional terms and the term challenging behavior. Therefore, challenging behavior was also interpreted and defined by each participant.
Participants were invited to comment on whether the findings were a reasonable explanation of their views to check the face validity of the individual repertory grids in a second meeting with the researcher. A team feedback session was also held to check for face validity at group level.
Individual and group-level repertory grid data were analysed using a statistical analysis package (Rep IV Research Version 1.12; Gaines & Shaw, 2005). The analyses involve a hierarchical cluster analysis (Shaw, 1980) to demonstrate the clustering of constructs and elements within each grid, a principal component analysis to show interrelationships between the elements and constructs in each grid (see Table 1) (also represented visually in a diagram for participants), and SocioNet analysis (Figure 1) to show the degree of commonality of construing, and hence mutual comprehensibility or otherwise, across the participants by examining the similarity of the structure of the participants' element clusters. The standard default settings for RepGrid IV were used throughout the analysis with the individual interpretation of the repertory grids being based on the work of Jankowicz (2004). A formal qualitative analysis of the data was not undertaken, but the idiosyncratic construct labels and quotations from the participants as they completed the repertory grid were used to put the repertory grid findings in context and to enable individual differences and common themes to be explored.
Table 1 summarizes the principal components from the 14 individual analyses that account for more than 80% of the variance in the individual grid of each participant (represented by P). Only one participant had a grid with a single component that accounted for this amount of variance, indicating that this person's grid exhibited low cognitive complexity (Bell, 2004) on this topic (P3). Nine participants had three principal components emerge from their repertory grids, suggesting high cognitive complexity of their construing of people with intellectual disability and beliefs about behavior that challenges.
The labeling of the largest components suggested that a range of salient constructs were elicited, including those relating to the presence of behavior that challenges, intellectual disability, diagnoses, personal interests, the causes of behavior problems, level of insight, personality, and relationships.
Internal versus external
The clients whose challenging behavior was caused by internal and external causes were not construed in systematically different ways in the repertory grids. When exploring how they construed the elements in the follow-up discussion, the participants did appear to ascribe different meanings to internal causes (being socially withdrawn, the presence of autism, experiencing trauma) and external causes of challenging behavior (inappropriate accommodation, responses of people around the client, clients' dissatisfaction with their living arrangements), but these were not always events that immediately preceded the behavior (e.g., past trauma).
In control versus not in control
Being in control of their behavior was associated with a range of constructs, the majority of which were seen more negatively; for example, miserable, quick tempered, purposely upsetting people, liking their own way, and making your job more stressful. In contrast, mental and physical health problems were thought to be the cause of uncontrollable behaviors, and staff tended to construe these clients as having higher levels of need. Behaviors were sometimes perceived as uncontrollable by the client if the carer could not identify any triggers. For example, a behavior was seen to be under the client's own control because she or he had responded well to a star chart and could be settled by the suggestion of an activity (P10) or, because if the client was told he or she could go tomorrow, a behavior would stop (P8). Although several participants construed in control of their behavior and not in control of their behavior as polar opposites, another participant (P5) stated that it was not possible per se for a client to be in control of their behavior.
Permanent versus temporary
People whose behavior was thought to be permanent, enduring, or unlikely to change were identified by a long history of exhibiting those behaviors or coming into services—the behavior being perceived as unchanging or the client being perceived to have a personality disorder. These clients were most frequently positioned closely to the challenging behavior constructs. However, P5 stated that it was not possible for a client's behavior to be permanent. It seemed that behaviors were perceived to be temporary if they changed in response to the environment, including through medical or psychosocial interventions. Thus, it seemed that the stability of behavior was a meaningful construct; although clients with temporary behavior problems were not consistently construed in the same way by participants.
Clients versus carers
The ideal client, the self, and the client without challenging behavior were placed close together by eight participants (along with the carer who struggles by five) at the preferred end of an important construct around behavior or functioning. However, two participants did not construe themselves as very different from the clients. In contrast, other participants appeared to use the self as a way to construe clients. For example, P3 said, “Ideally you like working with people who are like yourself don't you?” Another participant appeared to construe others in terms of her own personality traits and interests.
For the carer who struggled, most participants who selected a paid carer perceived them to be afraid and avoidant of behavior that challenges, which was seen as problematic. The ideal client was construed as not needing to come into an assessment and treatment unit. Several participants joked that the ideal client may be “boring,” being consistent with a marked preference for working with behavior that challenges, “My job would be very boring if they were all less demanding” (P10).
Clients were construed in terms of their life experiences (e.g., loss or trauma) by nine participants as well as in terms of mental health problems or medical/psychiatric diagnoses by six participants. Labels such as schizophrenia, personality disorder, and perhaps unsurprisingly within an intellectual disability service, autism were used as ways to explain behaviors.
Half the participants reported that negative experiences of carers or services had contributed to clients' difficulties, and the influence of family, friends, and other people on clients' well-being was also identified. For example, “I think we all need someone to lean on, every now and again … and if parents haven't shown a great interest in the development of their potentials and what have you, then they don't stand a cat in hell's chance out there” (P12), and “If someone's been labelled with, even with a minor learning disability, your normal mental health [service] won't take them on … and I just think sometimes I see them in the wrong environment … and it can't be too good for their progress” (P5).
Clients' insight into their own behavior also appeared to be relevant to six carers. However, some apparent inconsistencies also emerged. In particular, P5 reported that it was not possible for a client to be in control of challenging behavior, and the construct “knows they will cause harm—doesn't intend to do any harm” was closely associated with the element not in control of their challenging behavior. The severity of the client's disabilities was construed implicitly within four grids as associated with challenging behavior, and these were toward the less preferable pole of constructs.
The positives or rewards of working with behavior that challenges were also emphasized by some carers, such as job satisfaction from being able to work out what was wrong, being able to help the client, and being able to manage the challenge of their work: “I personally like people that are that little bit of a challenge, I think it gives me a job satisfaction that if you can talk them round, or you know, help them to understand or get, you know, do something, it gives you a personal challenge, and personal satisfaction” (P4). “In an ideal world, to have some of those behaviors still present, because you've got a bit of a chance then to dovetail what you're identifying to a future plan” (P12). However, some participants' grids and interviews revealed inconsistencies and ambivalence about these positive comments—for example, “Obviously on a one-to-one situation I prefer [a challenge], but obviously in the working day … you need the time to be able to go into the office and do things” (P1)—and by simultaneously placing these clients closest to the less preferable pole of constructs in the grids.
Analysis of Construing Across Participants
As the participants generated their own constructs, the present SocioNet analysis compared only the structural organization (clustering) of the elements across participants. In the SocioNet analysis (Figure 1), single-headed arrows signify unilateral understanding in the direction of the arrow (e.g., P14 pointing toward P1 indicates that P14 would understand P1 but not vice versa), and double-headed arrows signify mutually comprehensible understanding (e.g., P3 and P14 both understand each other), whereas an absence of arrows between any two participants indicates no common clustering of elements between them. Hence, a participant with no links to any other participants can be regarded as having a wholly distinctive way of construing his or her clients. To ensure the maximum amount of data was analyzed, P5 was excluded from this analysis as he only used eight of the elements.
Thus, the current SocioNet analysis indicated that P2 had no shared construing with any other participant. A further four participants (P1, P4, P9, P12) showed only one link to another participant. Participants P3 and P8 demonstrated the most reciprocal understanding with other participants, with more arrows pointing toward them than emanating from them. This suggested that their construct systems could be understood by others, better than P3 and P8 could understand other people's constructs. This is likely to be related to the narrow construing illustrated in their grids, with the most salient constructs being around level of everyday functioning or intellectual disability.
On the whole, the SocioNet confirmed the qualitative analyses, suggesting that there was a high degree of variability in the construct systems of the participants in the present study, despite the use of attributional dimensions within the elements.
As well as individual grids being validity checked, the general themes were fed back to the participating service. Staff agreed that the finding of diverse perceptions among the team appeared to be valid. They also reported that they agreed that there was a positive view toward working with behavior that challenges, and there appeared to be a culture of appearing fearless in the face of challenging behavior, with staff saying, “If we were scared, we'd be no use” and “If you're either strongly upset or scared, it's not the environment for you.” When asked whether this made it difficult to admit they were struggling at times, participants disagreed and said they were able to discuss this with colleagues.
The results suggested that the team did not hold a collective or stereotyped view of clients with behavior that challenges. Participants did appear to ascribe different meanings to the internal and external causes of challenging behavior, but those clients considered to have internal and external causes of their behavior were not actually construed in systematically different ways. Thus, it appeared that it was the construal of the whole person and her or his history that appeared to be the most relevant to understanding behavior in this staff team. Such a holistic perspective can be lost in questionnaire-based studies examining a priori models of staff responses to challenging behavior. Moreover, the differences between participants' construct systems would have been overlooked when using methodologies that examined the correlations between attribution and emotion ratings across the whole sample.
It is possible that appearing calm is part of successfully dealing with behavior that challenges (Duperouzel, 2008) and may serve to protect the self-concept as being good at this difficult job. The inconsistencies and ambivalence about these positive comments may be related to the importance of the carers' workload at that moment (e.g., Snow, Langdon, & Reynolds, 2007) and may not have been captured by questionnaires and hypothetical scenarios.
The present findings could possibly suggest that the participants were accustomed to challenging behavior because of their wealth of experience (e.g., Hastings, Tombs, Monzani, & Boulton, 2003) and their attitude of acceptance about the challenges of the role: “It's just, you know it's just part of the job. You just accept everybody's little ways and things and try and you know, sort of get on with it. … I don't really mind[;] it's just part of my job at the end of the day, it just comes with the job” (P4). However, the findings could also suggest carers were perhaps minimizing the negative experiences of caring for this client group because of the expectations within the service that carers should be nonjudgmental, like all of the clients, and be strong and fearless (e.g., Barnes, 1990; Rose & Cleary, 2007), or they could reflect a social desirability bias, as has been commonly reported in previous studies (Lucas et al., 2009; Rose & Rose, 2005).
Implications for Attribution Theory
The results suggested that attributions were one way among others for carers to understand people with intellectual disability and behavior that challenges and the finding that carers apparently held inconsistent and ambivalent views about the causes of behavior was consistent with personal construct theory (Kelly's “fragmentation corollary,” 1955). Kelly's “commonality corollary” states that differences in construing are to be expected, but that overlapping constructs allow people to reach common ground, and it can be noted that the current data lend support to the attribution models, in that clients presenting with challenging behavior were construed as being more in control of behavior that was permanent and related to internal causes.
The approach taken in the present study means that the specific constructs and themes that emerged will not be generalizable across all intellectual disability services or carers. In addition, it is acknowledged that the inclusion of broadly attributional dimensions within the elements to build on previous studies will have influenced the themes elicited, as will the supplying of the elements to the participants to some degree. Although this is a feature of some repertory grid studies, it does increase the possibility of increased commonality in the subsequent construing. However, the degree of observed heterogeneity in the current findings indicated that this was not a dominant factor. It is also possible that there was a social desirability bias within the findings (e.g., minimizing the negatives of behavior that challenges), although arguably this would have less impact than in questionnaires asking about intention to help.
The current study demonstrates the potential utility of repertory grid techniques as a means of readily examining the prevailing culture within services without losing the individual perspective and using average ratings. In this study, examination of the individual grids suggested that the prevailing culture was neither homogenized nor egalitarian, in that some staff may have marginalized on account of their different construing of and their behavior. This appeared to be confirmed by the SocioNet. Such individual differences could be helpful as different carers may develop better relationships with different clients. However, it also raised the question of how clinical decisions are made within the team and whether carers respond in inconsistent ways to behavior that challenges.
The findings suggested that training should take into account the social context within which the carers work and the impact of idiosyncratic perceptions on how training is applied. Beliefs about short-term reactive strategies that focus on harm prevention may be different from beliefs about longer-term proactive strategies that focus on addressing the causes of the behavior and improving quality of life (Hastings, 1997). Thus, carers' knowledge about what is a “correct” response may not translate into action (Oliver, Hall, Hales, & Head, 1996). Dunn, Clare, and Holland (2010) proposed that support workers made day-to-day “best interests” decisions within a framework of their own morals and values, consistent with the ideal client being construed as not needing to come into an assessment and treatment unit by some participants, perhaps reflecting carers' wishes for the clients to “live a life like ours” (Dunn et al., 2010). Permanent or enduring problems were linked by some carers to clients being perceived to have personality disorders, hitherto regarded, albeit inappropriately, as untreatable (National Institute for Mental Health in England, 2007), although it was not possible to determine exactly what participants understood by this term. Thus, it would seem important that staff teams are provided with up-to-date information about diagnostic terms such as personality disorders to foster optimism about change.
Behaviors were sometimes perceived as uncontrollable by the client if the carer could not identify any triggers. In contrast, some understanding by carers of the factors that led to behavior that was challenging was sometimes linked to the perception that the client was in control of her or his behavior. Given that controllable behaviors were perceived more negatively (consistent with Barrowclough, Gregg, & Tarrier, 2008), it would be important that carers observing and recording setting conditions or antecedents to problem behaviors are supported in the interpretation of such data.
The current method is arguably less susceptible to the social desirability biases than the use of self-report questionnaires. Moreover, through listening to the participants' quotes, it was possible to explore concerns about the social influences on participants' responses. For example, it was hypothesized that collective beliefs may exist in the service, which could be summarized as, “If you are scared of challenging behavior, you are bad at your job”; that carers should be nonjudgmental of all clients and behaviors; and that a theme emerged across some participants of staff wishing to present themselves as strong and fearless. This clearly has implications for carers' responses. On the one hand, it is possible that appearing calm is part of dealing with challenging behavior successfully (Duperouzel, 2008) and may serve to protect the self-concept as being good at this job. There was no evidence of stereotyping by the team, and the carers' perceptions of the person as a whole appeared to be more relevant to them than the perceived causes of the client's behavior problems. On the other hand, it raised the question of whether carers would feel able to voice fears of a particular behavior or their difficult feelings toward a client, and how this could influence their responses to clients. For example, it could lead to avoidance of clients (Bailey et al., 2006; Hastings, 2002). These issues suggested that team members may benefit from both clinical supervision, to which many nurses do not have access (Waskett, 2010), and facilitated discussions in which teams are facilitated to reflect on their different views in an atmosphere of mutual trust and respect, rather than trying to avoid disagreements (Onyett, 2007).
Recommendations for Future Research
Although not a direct test, the current findings provide some independent support for attributions about challenging behavior being important in how staff construe and act in the face of such behavior. In so doing, this demonstrates that methods other than the predominant vignette-based techniques are both feasible and valid for investigating staff views about challenging behavior. Moreover, the repertory grid technique as demonstrated in the present study would appear to have validity for investigating team and service culture in both an exploratory manner and as a means of evaluating staff training and supervision initiatives. In addition, there is very little, if any, research to date on carers' enjoyment of working with people with behavior that challenges, or the culture/attitude of being strong and fearless, which may deserve further study. Several authors have proposed that excitement attracts people to work in critical care and emergency nursing and medicine (Cronin & Cronin, 2006; McLarnon & Wise, 2003), and it may be of benefit to explore this concept in the field of challenging behavior.
The use of repertory grid techniques elicited a range of concepts used by carers to construe people and behavior. In addition, the analyses showed that only a minority of the participants shared a common construct system on this topic. Arguably, the attributional theory of helping behavior is not sufficient to explain staff responses in this field. Future research should acknowledge the subjective and diverse ways that carers understand people with intellectual disability and behavior that challenges and should pay attention to the context within which carers' cognitive, emotional, and behavioral responses emerge.
The authors would like to thank the staff who participated in the present study as without their support and commitment, this work would not have been possible.
Editor-in-Charge: Glenn T. Fujiura
Dougal Julian Hare (e-mail: email@example.com), University of Manchester, School of Psychological Sciences, Zochonis Building, Brunswick Street, Manchester, M13 9PL, United Kingdom; Marianne Durand, Manchester Learning Disability Partnership, Manchester, United Kingdom; Anja Wittkowski, University of Manchester, Manchester, United Kingdom; and Steve Hendy, Trafford Council, Sale, United Kingdom.